Heidelberg weekly meeting

03.410 (Physikalisches Institut, Heidelberg)


Physikalisches Institut, Heidelberg


We will have the meeting in HeiCONF: https://heiconf.uni-heidelberg.de/sch-nqu-dd2   Access Code: 858824

Vidyo will be used as a back-up. Room extension 10823320 or use this link.


Tentative minute taker list at https://docs.google.com/spreadsheets/d/15oo_a_1WPxPBOtlH0LDEKG1128y1O2lfswpUAvPCQOY/edit#gid=0

    • 14:00 14:45
      Round Table 45m

      Reports of work progress

      Speakers: Adrian Herkert (Physikalisches Institut, Uni Heidelberg), Alena Weber (Universität Heidelberg), Andre Schoening (University Heidelberg, Institute of Physiscs), Annie Meneses Gonzalez (PhD Student), Benjamin Weinläder (PI Heidelberg), David Maximilian Immig (Universität Heidelberg), Dohun Kim (Physikalisches Institut University of Heidelberg), Frank Meier Aeschbacher (Universität Heidelberg), Heiko Augustin (PI Uni Heidelberg), Lars Noehte (PI Heidelberg), Luigi Vigani (University of Oxford), Lukas Mandok (Physikalisches Institut Heidelberg), Sebastian Dittmeier (Physikalisches Institut, Universität Heidelberg), Sebastian Preuß (PI Uni Heidelberg), Thomas Theodor Rudzki (Physikalisches Institut Heidelberg)

      Date: 16.07.2020

      Minute Taker: Sebastian Preuß

      Present: Andre, Heiko, Alena, Alexander, David, Dohun, Jan, Konstantin, Lukas, Marius, Annie, Sebastian Preuß,


      Lukas: Presentation about his measurements (Characterization of ATLASPix 3)

      • reiterate measurements next week for testbeam preparation

      Dohun: Presentation (ATLASPix3 time resolution scanning)

      • Discussion on whether measurements are too noisy → unclear whether conclusions can be properly drawn

      • David suggests checking different VMinus range

      • VN: Best values 16,18 (decimal)

      • VNComp: Best value 17 (hex)

      • VNFB: Best value 17 (dec)

      • Andre: How much data was taken? → 10 runs (of ~5 min each) per configuration → same amount of statistics per setting

      • Andre: Measurements too noisy? → Dohun: maybe configurations are unstable, maybe

      • best time resolution (best config): 5.9ns fully corrected, 9ns uncorrected (VN: 10, VNComp: 17, VNFB: a (all hex), minTh 57mV)

      • Andre: Very small differences, maybe over-accounted some conclusions due to high statistical fluctuations → more study needed before definite conclusions are drawn

      • Further study: Does VNFB depend on VN? Scan digital part: VNBiasRec, VPBiasRec; find out how to decrease threshold => test best configuration at PSI

      • Andre: How does time resolution depend on threshold? → More corrections necessary due to changed ToT


      Round table:

      Thomas (on the train, only text comment): Si heater mockup is foreseen to be completely ready end of September (to be tested then at Brugg). Testbeam crew is kindly asked to pick up the mini cage from Brugg during the August campaign



      • TDAC configuration of MuPix 10, old config takes 6 secs via SPI (→ would take 2 hrs for full tuning per chip)

      • Implemented & debugged faster SPI with luigi → Full chip config now possible in 1 minute

      • Allow Marius further study now

      • Andre: Importance? How many noisy pixels were encountered at e.g. DESY? → Quick check with Fe-55 showed about 3-5 pixels going hot very fast

      • Andre: Plan for PSI? Is a chip always fully tuned? → For now only global config, sometimes masked out single pixels

      • Wrote function to zero out all configs, can/should be run once after startup (takes only 3secs though)

      • Can also switch all pixels on/off at the same time quickly

      • Bitbucket-Issue to be looked at by David or Luigi for further work


      • Tuedsay: MuPix 10 PCB arrived, will be bonded tomorrow (17.07.2020), will be finished mid to end of next week

      • Firmware changes for MuPix 10 finished

      • Worked on documentation


      • Workshop ordered 42 new Inserts (should arrive in ~2 weeks)

      • LTU (six-chip-)flex: New version including SPI line

        • power delivery and routing should work; chip 6 was

        • Question: Where exactly do the SPI lines go?

        • Flex should work for current SPI configuration setup

        • Interposer flex: Second version for tests with tap-bonds?

      • Thomas: Use 6-chip-flex without interposer flex? → Easier solution, just use rigid PCB

      • Thomas: Are any sizes changed? → No

      • Andre: Maybe simulate power distribution? → Alex should send positions of cuts for new simulation to calculate power drop

      • Maybe further discussion when simulations are done

      • Probe card arrived → Still has to be tested (early tests indicate it is working well though)

      • Oxford completed their ladder design → Very different from the one done in Heidelberg

        • They route all data lines to the very bottom of the flex

        • On interposer: all data lines in one corner, rest reserved for power

        • Andre: What is the further plan? → Will discuss with Luigi

        • Proper agreement and internal review necessary before sending anything to LTU

      • Single ladder design is finished

      • Do we want the sync reset signal or can we do without it? Heiko: If we do not use mu3e slowcontrol we do (i.e.: yes)→ Alex needs one more day for completing it

      • Do we have USB-C cables for testing? Interesting because the plug is symmetrical, the internal routing is not

        • USB-C protocol takes care of that to match pairs that are connected, but we use it out of specification

        • Jan suggests labeling cables in order to plug them in the correct way

        • Andre: Design pinout such that nothing is damaged when the cable is accidentally plugged in the wrong way


      • Looked into offline sorter, created control histograms

        • Link-link, chip-chip, … delays can be controlled

      • Will continue with MuPix 10 Analysis

        • Currently seeing efficiency decrease in Matrix C, but unclear if this is due to offline sorting (testbeam data (DESY 202006/07) indicates this)


      • Presentation was given (see above)


      • Information regarding talk next week will be sent

      • Inserts: Ralf is happy with the scheme, says it can be bonded

        • David: External power connection? → Probably not needed

        • → David: Direct connection is better than motherboard readout → Jan will design better testpoint

        • Further discussion will take place in Mattermost


      • Finished first implementation for pattern recognition bank

      • Will be able to run whole analysis soon

      • Looking at solution on how to store the lookup table, will then start filling it

      • After that: Test with background data and test whether it will be rejected as intended

      • Heiko: Had MuPix 10 running in the lab without source (→ only cosmics), saw 90% single clusters, but a lot of delta-events. Are these considered (probably beyond the scope of Konstantin’s thesis)? → Part of actual reconstruction => not part of thesis

        • Maybe someone has to take a deeper look at this

        • Andre: Are we sure these are actually muons? Heiko: Quite sure, yes

        • Andre: What are the energies of the particles Heiko saw (Only interested in high energy in cosmic trigger)? Heiko: Probably needs more looking into

        • Andre: Maybe measure this when modules are available, or do stacked measurements

        • Konstantin: How large are these clusters? Heiko: Highly asymmetrical, 4-10px, but not important for trigger because of superpixels


      • Talk given – see above


      • Built MuPix10 setup

      • Adjusted measurement software with Luigi

      • Can investigate now with software written by Heiko and Luigi (see above)

      • Injection only works for 2/3 of pixels simultaneously, will be investigated

      • Will implement some corrections in the software due to effects seen

      • Hopefully will be able to do offline tuning by the end of next week → maybe share results?


      • ROOT on tachyon is not working? Apparently nobody else has experienced this, but it is unclear who actually uses it


      • Work on plots, less progress due to advanced lab course.


      • Was at PSI for preparing irradiation setup

      • Frank will be back in 1,5 weeks and if Lukas Gerritzen still needs help will then be able to provide it

      • Silicon heater ladders should be ready by the end of this month → There will be measurements taking place in september whether cooling is still sufficient

      • Silicon heater programme will probably be finished then

      • Testbeam-crew should take wind-tunnel setup with them on the way back → reminder will be sent


      • Magnet will be delivered to PSI starting next week

      • Change in kryogenics plannig: Final testing in sheffield will not be taking place – assembly has taken place in London

      • Final test will take place in London next week; if eeverything works: Ship it on friday (24.07.2020) or monday (27.07.2020) → will be shipped by truck, which has to be on a ferry (too large for tunnel), unclear how the delivery will work out, but should take ~2 days to psi


      The main meeting ends, some leave; testbeam preparation/planning for the rest of the year is discussed.

      Upcoming PSI testbeam preparation:

      • David: Nothing needs to be specifically ordered anymore

      • Andre: Will send information regarding corona safety measures

      • David: Institute bus not here, will not be lent for two weeks → Need to take out one at e.g. Europcar → Discussion on when to rent and when to give it back

      • David: Frank is on vacation, we have a new contact to talk to regarding going to higher energies, but should do a coarse test beforehand

        • Andre: Does rate go down at higher energies?

      • David: VNCompPix scan should be interesting (compare to old MuPix8 measurements – saw interesting effects)

      • Andre: What are the plans for ATLASPix3 after all scans are done?

        • David: Do full ATLASPix3 telescope or MuPix 10 telescope with AP3 DUT

          • Is quite advanced (Heiko says it’s not if properly tested in the lab), but we only have two working ATLASPix3 sensors (→ maybe re-bond stuff?)

          • David has no time to test a telescope, Heiko proposes to test ATLASPix3 in a MuPix10 telescope, this is maybe even easier in the analysis because of ATLASPix3’s timewalk characteristic

          • Andre: Have working Mupix 10 telescope from DESY as starting point, if it fails for ATLASPix3 we can just take MuPix 10 data instead

          • Heiko: Testing should start in the lab ; David suggests checking the new offline sorter with ATLASPix3 data then

        • Conclusion: We will try to set up an ATLASPix3 in a MuPix 10 telescope, will be tested in Heidelberg next week (20.07.-26.07.2020)

      • Regarding FPGAs:

        • For this year: No new STRATIX boards, no time for new developments

        • Sebastian Dittmeier said that recompiling the code for ARIA5 works

        • Benjamin might take over ARIA5 boards when he starts in August


      October testbeam preparation:

      • Heiko: Low participation from LHCb, only Vadim and Jan , rest is our crew

      • Heiko: How much preparation do we need?

        • Andre: Annie and Benjamin need to do preparation in the lab

        • David: One week per sensor to get it running and full optimisation estimated

        • Heiko: Invite Vadim and Jan to Heidelberg one month in advance?

      • Both Telepix designs are new and in some way non-standard

      • David suggests some characterisation should take place by LHCb people

        • Andre: First round we have to do it ourselves; we know our procedures and it’s feasible for us to do it. Familiarizing external people would take too long

        • Andre: Annie needs to do some work on these sensors anyways

        • Jan: Skeptical because of travel restrictions

      • LHCb wants to look into analysis; David: useful, because having a setup with sensors with different pixel sizes is maybe a hard task

        • Andre: Implementation should be possible, there’s plenty of time for analysis after data-taking, analysis is out of our (main) responsibility though

      • Andre: Our interest: Get CMOS comparators running – this is our best circuitry

        • Second step: Get CMOS amplifiers running

        • Ivan says that for smaller pixel sizes NMOS has advantages

        • => Can already learn a lot from chips “1” and “3” (compare slide uploaded by Andre)

      • David: Maybe two people is not enough, but none of the existing students are available

        • Andre: New students would need to be introduced to the project → Benni and Annie

      • Andre: Chip “2”: Should work well, because of our experience with NMOS and CMOS, but complicated pixel readout scheme

        • Heiko: Define two chips for analysis, split into sub-matrices and interpret the data from one physical chip as two logical chips in the analysis software

      • Andre: Ivan said someone from KIT will also participate → maybe discuss whether this person can also help out in the lab.

      • Jan: Which chips will be tested? Reply by Andre:

        • Chip “5” will definitely not be tested

        • 7 & 8 probably too ambitious → also crossed out

        • Week one: Chips “1” and “3”

        • Week two: “2” and “4” seems realistic, “6” depends on progress made with the other four

        • Jan: Chips are small → Hard to get statistics, estimates 1.5 – 2 chips per week

        • Andre: Don’t spend too much tome on “4” and “6” in preparation; if one works, the others should too

      • Heiko: If chip delivery works out as expected, we should have > 1 month of preparation time

        • David: September testbeam takes away manpower. But: We need this testbeam after DAC-optimisation in the Lab, cannot skip it over the upcoming PSI testbeam

        • Andre: It is still unclear who’s going in September, but maybe it’s just one week because other mu3e groups will be going as well. 1-2 people should probably be enough

      • Andre: Regarding LHCb presentation: Would be a good success if we get good results with the little preparation time we had. Will probably apply for testbeam time in January/February 2021 to do further tests if everything works well

      • Andre: Do we put a chip in the MIMOSA telescope in October testbeam?

        • Annie: has experience and could do it

        • General discussion on whether it is too ambitious. Conclusion: Discuss in further weeks what is feasible

        • Andre: Can Lennart help? → Heiko: Work on two separate setups creates problems

          • Can Lennart do some work in Heidelberg?

      • Andre: Get Ivan to Heidelberg because documentation is lacking


      After 3:40 hrs the meeting ends

    • 14:45 14:55
    • 14:55 15:15
      Run2020 20m
      Speaker: Andre Schoening (University Heidelberg, Institute of Physiscs)